CO2 leakage in shallow aquifers: A benchmark modeling study of CO2 gas evolution in heterogeneous porous media

https://doi.org/10.1016/j.ijggc.2015.04.017Get rights and content

Highlights

  • Simulated intermediate scale column experiments of CO2 gas exsolution in heterogeneous porous media.

  • Developed and validated an expression to predict the critical pressure at which free phase CO2 gas forms.

  • Our simulations capture experimental observations, including CO2 gas accumulation zones in heterogeneous sands.

Abstract

The physicochemical processes associated with CO2 leakage into shallow aquifer systems are complex and span multiple spatial and time scales. Continuum-scale numerical models that faithfully represent the underlying pore-scale physics are required to predict the long-term behavior and aid in risk analysis regarding regulatory and management decisions. This study focuses on benchmarking the numerical simulator, FEHM, with intermediate-scale column experiments of CO2 gas evolution in homogeneous and heterogeneous sand configurations. Inverse modeling was conducted to calibrate model parameters and determine model sensitivity to the observed steady-state saturation profiles. It is shown that FEHM is a powerful tool that is capable of capturing the experimentally observed outflow rates and saturation profiles. Moreover, FEHM captures the transition from single- to multi-phase flow and CO2 gas accumulation at interfaces separating sands. We also derive a simple expression, based on Darcy's law, for the pressure at which CO2 free phase gas is observed and show that it reliably predicts the location at which single-phase flow transitions to multi-phase flow.

Introduction

Long-term storage of CO2 in geologic formations is one of the most promising technologies aimed at significantly reducing anthropogenic carbon emissions (Pacala and Socolow, 2004). Substantial reductions in emissions will require large quantities of injected CO2, resulting in CO2 plumes that could occupy areas of 100 km2, or more, which will likely intersect leakage pathways out of intended storage formations (Pruess, 2008). Moreover, the injected CO2 will displace large quantities of brine, some of which is CO2-saturated. Thus, the potential for leaked CO2 or displaced brine to migrate into shallow groundwater aquifers is a concern that must be addressed. This requires knowledge of the physicochemical processes associated with CO2 flow and transport, as well as tools (i.e., numerical models) that can capture these processes and predict long-term behavior.

Possible leakage scenarios and their associated physicochemical processes have been described in the literature (e.g. Pruess, 2008, Plampin et al., 2014a). In this work, we focus on CO2 gas evolution through shallow aquifers, where evolution refers to the combined processes of CO2 gas exsolution, accumulation, and free phase migration. These processes are complex and span multiple spatial and time scales. In this leakage scenario CO2 is initially dissolved in displaced brine that migrates through the aquifer where CO2 bubbles can exsolve out of solution due, in part, to decreases in pore pressure. These bubbles may either rise individually through the pore space due to buoyant forces (e.g. Corapcioglu et al., 2004, Oldenburg and Lewicki, 2006) or expand while trapped, forming gas clusters that invade pores through invasion-percolation (e.g. Li and Yortsos, 1995, Li and Yortsos, 1995, Dominguez et al., 2000). As more gas exsolves, the bubbles and gas clusters coalesce to form a continuous CO2 gas phase capable of migrating throughout the aquifer. The minimum gas saturation at which a continuous free phase gas forms and becomes mobile is known as the critical gas saturation, Sgc. This parameter is important for describing CO2 gas evolution with continuum models since they rely on the concept of a representative elementary volume (REV), and therefore cannot account for bubble nucleation and exsolution dynamics at the pore scale. Although Sgc is a parameter of significant importance, ambiguity still exists regarding its definition, and reported experimental values vary widely (e.g. Du and Yortsos, 1999, Sheng et al., 1999, Tsimpanogiannis and Yortsos, 2004, and references therein).

Heterogeneities are ubiquitous in the subsurface and are expected to affect the evolution of CO2 gas. However, most laboratory studies have focused on bubble and cluster dynamics in homogeneous micromodels (e.g. Li and Yortsos, 1995, Du and Yortsos, 1999, Tsimpanogiannis and Yortsos, 2004, Zuo et al., 2013) or free phase migration in homogeneous sands and rocks (e.g. Fry et al., 1997, Enouy et al., 2011, Zuo et al., 2012, Krevor et al., 2012, Falta et al., 2013, Zuo and Benson, 2014). Recently, Sakaki et al. (2013) and Plampin et al., 2014a, Plampin et al., 2014b conducted intermediate-scale CO2 gas evolution experiments in homogeneous and heterogeneous sand configurations. They showed that simple heterogeneities had significant affects on CO2 gas evolution, especially at interfaces between coarse and fine sands where they observed increased accumulation of CO2 in coarse sands when located underneath a layer of fine sand. In addition, they observed values for Sgc ranging from 0.2 to 0.4 within their sands. Although several numerical simulators have been developed specifically for CO2 sequestration and leakage applications (e.g. Class et al., 2009), benchmark studies involving comparisons with laboratory experiments are scarce. In fact, we are aware of only one such study conducted by Enouy et al. (2011), in which they compared their model to CO2 gas evolution column experiments in homogeneous glass beads and sands. In their model, they assumed rate-limited mass transfer of CO2 from aqueous to gas phase and showed that advection of the gas phase is adequately modeled with the classical extension of Darcy's law for multi-phase flow and conventional constitutive relationships. However, recent experimental studies have suggested that non-conventional relative permeability relationships may be required to accurately represent CO2 gas evolution in porous media (Krevor et al., 2012, Falta et al., 2013, Zuo and Benson, 2014). In fact, Zuo and Benson (2014) conducted microtomography experiments comparing CO2 gas exsolution and immiscible displacement and concluded that the traditional model parameters and constitutive relationships are process-dependent. Similar observations have been reported in oil- and gas–water systems (e.g. Fishlock et al., 1988, Grattoni et al., 2001). Thus, there is a need to benchmark existing numerical simulators with experimental observations and test conventional parameterization methods for these models, especially in heterogeneous media.

Accordingly, the objective of this work is to asses how well the numerical simulator, FEHM (https://fehm.lanl.gov), can capture CO2 gas evolution in the homogeneous and heterogeneous intermediate-scale CO2 leakage experiments conducted by Plampin et al., 2014a, Plampin et al., 2014b. In Section 2, we provide a brief summary of the experiments followed by a detailed description of FEHM and the simulations. FEHM differs from the model employed by Enouy et al. (2011) in that equilibrium mass transfer of CO2 from aqueous phase to gas phase is assumed, which is calculated using a variable-switching method Forsyth and Sammon (1984). Based on the lack of model parameter data for CO2 gas and water, we conducted inverse modeling using PEST (http://www.pesthomepage.org) to estimate suitable parameter values and determine model sensitivity. In addition, we developed a simple analytic expression, based on single-phase flow theory, to estimate the critical pressure at which CO2 can exsolve out of solution to form free phase CO2 gas. Plampin et al. (2014a, Section 3.1 and Fig. 4) developed a similar expression based on Henry's law and hydrostatic conditions within the column. However, we have extended that relationship to account for the additional pressure drop due to flow, which is applicable to both homogeneous and heterogeneous systems. The results are presented in Section 3, where we compare experimental and simulated outflow rates and steady-state saturation profiles. In addition, we show that there is agreement between experiments, numerical simulations, and theory for the depth at which CO2 evolution is observed and that FEHM captures CO2 gas accumulation at interfaces between different sands. We conclude with a summary and discussion regarding the implications of this work for shallow aquifer systems in Section 4.

Section snippets

Experimental system

Here, we summarize the most important features of the experiments, complete details can be found in Plampin et al., 2014a, Plampin et al., 2014b. Fig. 1 illustrates the experimental systems and sand configurations, which consisted of two different columns. The first column was 4.5 m tall by 0.06 m diameter (hereafter called the long column) packed with two homogeneous and three heterogeneous sand configurations. The second column was 1.68 m tall by 0.06 m diameter (hereafter called the short

PEST analysis

The model parameters that are most important for these experiments are the residual wetting phase saturation, Swr, the critical gas saturation, Sgc, the pore-size distribution parameter, λ, in the Brooks–Corey relative permeability equation, and the two parameters, α (the inverse of the entry pressure) and n (the pore-size distribution), in the van Genuchten capillary pressure – saturation relationship. Conventionally, these parameters are measured from drainage or imbibition experiments within

Discussion and conclusions

In this work, we conducted FEHM simulations of intermediate-scale CO2 leakage experiments in homogeneous and heterogeneous sand configurations conducted by Plampin et al., 2014a, Plampin et al., 2014b. The experiments were designed to investigate the combined processes of CO2 gas exsolution, free phase migration and accumulation. The long-term affect of these processes within shallow aquifers constitutes a leakage scenario that must be understood to ensure CO2 sequestration operations do not

Acknowledgments

This work was funded by U.S. Department of Energy's CO2 Sequestration R&D program and managed by National Energy Technology Laboratory. The experiments were performed at the Center for experimental study of subsurface environmental processes (CESEP) at Colorado School of Mines.

References (40)

  • H. Class et al.

    A benchmark study on problems related to CO2 storage in geologic formations

    Comput. Geosci.

    (2009)
  • M.Y. Corapcioglu et al.

    Rise velocity of an air bubble in porous media: theoretical studies

    Water Resour. Res.

    (2004)
  • A.T. Corey

    The interrelation between gas and oil relative permeabilities

    Prod. Mon.

    (1954)
  • C. Du et al.

    A numerical study of the critical gas saturation in a porous medium

    Trans. Porous Med.

    (1999)
  • Z. Duan et al.

    Densities of the CO2–H2O and CO2–H2O–NaCl systems up to 647 k and 100 mPa

    Energy Fuels

    (2008)
  • R.W. Falta et al.

    Migration of exsolved CO2 following depressurization of saturated brines

    Greenh. Gases: Sci. Technol.

    (2013)
  • T. Fishlock et al.

    Experimental studies on the waterflood residual gas saturation and its production by blowdown

    SPE Reserv. Eng.

    (1988)
  • P. Forsyth et al.

    Gas phase appearance and disappearance in fully implicit black oil simulation

    Soc. Pet. Eng. J.

    (1984)
  • V. Fry et al.

    Experimental investigations for trapping oxygen gas in saturated porous media for in situ bioremediation

    Water Resour. Res.

    (1997)
  • J.E. Garcia

    Density of Aqueous Solutions of CO2

    (2001)
  • Cited by (0)

    View full text